WO2015084154A1

WO2015084154A1 - A system and method for locating a mobile device

Info

Publication number: WO2015084154A1
Application number: PCT/MY2014/000195
Authority: WO
Inventors: Keeratpal SINGH; Fazreen Binti Mohd Apan Affiza
Original assignee: Mimos Berhad
Priority date: 2013-12-04
Filing date: 2014-06-26
Publication date: 2015-06-11

Abstract

The present invention relates to a system and method for locating a mobile device. More particularly, the invention relates to a system and method for locating a mobile device by using heterogeneous neural network. The system comprises of at least one mobile device (10), a mobile enterprise platform (20) and at least one telecommunication service provider (30). The mobile enterprise platform (20) further includes a location server (21), a database (22), a Phone Listener Application (23), a Collector Application (24) and a Quality of Positioning (QOP) Mapper (25).

Description

A SYSTEM AND METHOD FOR LOCATING A MOBILE DEVICE

FIELD OF INVENTION

The present invention relates to a system and method for locating a mobile device. More particularly, the present invention relates to a system and method for locating a mobile device by using heterogeneous neural network.

BACKGROUND OF THE INVENTION

A mobile device is typically provided with a location-based service application wherein the application may include features to find directions or to detect location of the mobile device. There are several systems and methods developed to assist the operation of such application.

An example of such system and method is disclosed in US Publication No. 2013/0072215 A1 , whereby the location prediction device provides a location prediction of a mobile device based on an adaptively compiled visitation history. The location prediction may be performed without the use of large amounts of system resources. The location prediction may be used in conjunction with any mobile device application known in the art. The prior art includes the use of global positioning system (GPS) to provide the location for the mobile device.

Another example is disclosed in US 2009/0197612 A1 whereby a mobile device includes a location determination system for determining a location of mobile device and a system for sending location information of the mobile device. The location determination system includes a GPS receiver. The mobile device can comprise a system for displaying the location of the second mobile device and/or perform a task of a non-location specific application, such as a music player.

Based on the examples above, the system and device implement GPS to assist the location-based service application. GPS requires a clear admission to the satellites to provide the location tracking and maps on its devices. The GPS is not available if the mobile device is in a closed area such as inside a building or a tunnel. Thus, the operation of such location-based service application is limited. Hence, there is a need to improve the system and method for the application to address such drawbacks. SUMMARY OF INVENTION

The present invention provides a system for locating a mobile device. The system comprises of at least a mobile device (10), a mobile enterprise platform (20), at least one telecommunication service provider (30). Moreover, the mobile enterprise platform (20) is characterised in that it further includes a location server (21) for locating the mobile device (10) in different types of networks; a database (22) to store collected data including Cell Identification (ID), receive signal levels (RXLs) and the location coordinates of the mobile device (10); a Phone Listener Application (23) to respond to request from the location server (21) and collect RXL of the mobile device (10); a Collector Application to collect RXLs and configuration setting sent by the location server (21); and a Quality of Positioning (QOP) Mapper (25) to map the quality of positioning (QOP) level with the type of network required. Preferably, the Phone Listener Application (23) includes a location engine application to provide estimation without involving the server of the location server (21).

The present invention also provides a method for locating a mobile device (10). The method is characterised by the steps of collecting and processing data including Cell Identification (ID), signal levels (RXLs) and location coordinates of the mobile device (10); performing a neural network training pattern and optimization; performing a location request with a Quality of Positioning (QOP) Mapper (25); performing a location request with the telecommunication service provider (30); searching for location server (21); operating the mobile device (10) with the Phone

Listener Application (23); and performing a server search call if there is no internet connection and the coordinates of a last location of the mobile device (10) is not available in the database (22). Preferably, the step of collecting and processing data of the mobile device

(10) includes sending a command to the Collector Application (24) for collecting RXLs and configuration setting by the location server (21); changing network settings by turning on the GPS; recording Cell ID and RXLs in a Training Table by the Collector Application (24); sending the recorded Cell ID and RXLs to the location server (21) within pre-configured time interval; and turning off the GPS by the Collector Application (24) when a maximum sample size (N) set by the location server (21) is achieved.

Preferably, the step of processing a neural network training pattern and optimization includes arranging the collected data based on Location Area Code (LAC); arranging the RXLs based on the Cell IDs, network services, GPS coordinates, time and date; training neural network algorithm to derive a pattern for a particular location area based on a predefined percentage of samples of the collected data; testing the neural network algorithm using a predefined percentage of samples of the collected data; determining a smooth parameter (σ) based on a General Regression Neural Network (GRNN); plotting a Cumulative Distribution Function (CDF) graph by using the collected data; measuring the distance between two points in the CDF graph to find the difference of the location coordinates; comparing the smooth parameter (σ) with a benchmark; calculating a new correction factor (p) if the smooth parameter (σ) is within the benchmark or the difference (Diff **) between the first iteration value and the second iteration value is too small; and storing the value of p in the database (22).

Preferably, the step of performing a location request with the QOP Mapper (25) includes mapping a Quality of Positioning (QOP) level with the type of network services required by the QOP Mapper (25); sending the configuration setting to change the network service based on the QOP level required and the location request to the Phone Listener Application (23) by the location server (21); measuring the Cell ID and RXLs based on the QOP level required by the Phone Listener Application (23); translating the Wi-Fi coordinates to latitude and longitude coordinates in a map by the location server (21). The step of measuring the Cell ID and RXLs when the QOP level is low suitably includes obtaining the Cell ID, RXLs and Wi-Fi information; and translating the Cell ID coordinates to a location coordinates by the location server (21), wherein if the mobile device (10) is in indoor area, the Wi-Fi coordinates is used as the proximity location. The step of measuring the Cell ID and RXLs when the QOP level is medium suitably includes obtaining the RXLs of that particular network service and Wi-Fi network service; sending the RXLs of that particular network service and Wi-Fi network service to the location server (21); matching the RXLs of that particular network service and Wi-Fi network service with the stored RXLs using a Statistical Method and Free Space Propagation Loss Formula, wherein the RXLs are converted into distance; and triangulating at least three base stations coordinates to estimate the location coordinates. The step of measuring the Cell ID and RXLs when the QOP level is high suitably includes obtaining Cell IDs and RXLs of the networks setting by changing the network services; sending the Cell IDs and RXLs of the networks setting to the location server (21); loading the value of p, and GPS coordinates of the Cell IDs that have been processed by the neural based on the collected data to estimate the current location of the mobile device (10); and estimating the current location by comparing the new Cell IDs or RXLs with the trained Cell IDs or RXLs.

Preferably, the step of performing a location request with the telecommunication service provider (30) includes sending the enterprise phone number of the mobile device (10) to the telecommunication service provider (30) by the location server (21); searching for the mobile device (10) by using a Mobile Location Centre by the telecommunication service provider (30); sending the coordinates of the mobile device (10) to the location server (21) if the SIM card is installed in the mobile device (10); sending the last active location of the mobile device (10) by the telecommunication service provider (30); and translating the location coordinates to latitude and longitude coordinates in a map by the location server (21).

Preferably, the step of operating the mobile device (10) with the Phone Listener Application (23) includes the steps of monitoring battery and connectivity status of the mobile device (10) by the Phone Listener Application (23); comparing the stored Cell IDs with the Wi-Fi IDs within the enterprise or comparing the stored Cell IDs with frequent travelled areas if there is no internet connection; locking the mobile device (10) within a pre-configured time interval if there is internet connection, Cell ID and Wi-Fi IDs; estimating the coordinates from the heterogeneous networks of RXLs by using any QOP level by the location server (21); and sending the coordinates of the mobile device (10) to the administrator if the Phone Listener Application (23) in the mobile device (10) is not able to do the server search.

Preferably, the step of searching the location server (21) includes sending a message or performing a voice call by the location server (21) if the location server (21) detects no active communication from the last location of the mobile device (10) within a preconfigured time interval; intercepting the message or voice call from the enterprise number and reading request command with high QOP level by the Phone Listener Application (23); obtaining the Cell IDs and RXLs of the networks settings by changing networks; sending the enterprise message number to the location server (21), wherein the enterprise message number is a pre-stored phone number connected to the enterprise location server; reading the messages by the location server (21), wherein the messages are extracted to proceed with location search using the General Regression Neural Network for the current Location Area Code (LAC); determining the proximity coordinates of the mobile device (10) by using the Cell ID coordinates if there is no trained data for the LAC; and translating the location coordinates to latitude and longitude coordinates in a map by the location server (21).

Preferably, the step of performing a server search call if there is no internet connection and the coordinates of the last location of the mobile device (10) is not available in the database (22) includes performing a voice call by the location server (21) if the location server (21) detects no active communication from the last location of the mobile device (10) within a pre-configured time interval; playing a recording to request the returning of the mobile device (10) to the address sent by the location server (21) via message if the voice call is answered; intercepting the voice call from the enterprise number by the Phone Listener Application (23); identifying the location of the mobile device (10) based on International Mobile Station Equipment Identity (IMEI) of the mobile device (10) by the telecommunication service provider (30) if the phone call or message of the mobile device (10) provides no response to the location server (21); obtaining the Cell IDs and RXLs of the network settings by changing networks if the GPS cannot be activated to provide location coordinates; sending the enterprise message number to the location server (21); reading the messages by the location server (21) if the message is successfully sent, wherein the text messages are extracted to proceed with location search using the General Regression Neural Network for the current LAC; determining whether there is trained data in the Neural Network Lite if the message is not successfully sent; providing the Cell ID, LAC, IMEI and RXLs information if there is no trained data in the Neural Network Lite; loading the trained stored data of the Neural Network Lite, value of p and stored GPS coordinates of trained Cell IDs and RXLs from the LAC and the Cell IDs if there is trained data in the Neural Network Lite; predicting the location coordinates by the Neural Network Lite if the GPS is not activated; and sending the Cell IDs and IMEI number of the mobile device (10) to an emergency number.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

FIG. 1 illustrates a system for locating a mobile device (10) according to an embodiment of the present invention.

FIGS. 2(a-h) illustrate flowcharts of a method for locating a mobile device (10) according to an embodiment of the present invention. DESCRIPTION OF THE PREFFERED EMBODIMENT

A preferred embodiment of the present invention will be described herein below with reference to the accompanying drawings. In the following description, well known functions or constructions are not described in detail since they would obscure the description with unnecessary detail.

FIG. 1 illustrates a system for locating a mobile device (10) according to an embodiment of the present invention. The system comprises of at least one mobile device (10), a mobile enterprise platform (20) and at least one telecommunication service provider (30). The mobile enterprise platform (20) further includes a location server (21), a database (22), a Phone Listener Application (23), a Collector Application (24) and a Quality of Positioning (QOP) Mapper (25).

The location server (21) is used for locating the mobile device (10) in different types of network services such as Global System for Mobile Communications (GSM), Global Positioning System (GPS), Third Generation telecommunication network (3G), High-Speed Downlink Packet Access (HSDPA) and Wi-Fi. The location server (21) communicates with the mobile device ( 0) via mobile enterprise platform (20). The database (22) stores relevant collected data such as Cell Identification (ID), receive signal levels (RXLs) and the location coordinates of the mobile device (10). The Phone Listener Application (23) is a standby application which responds to request from the location server (21) and collects RXL of the mobile device (10). The Phone Listener Application (23) functions with a light location engine neural network application. Light location engine does not involve comprehensive computations which the location server (21) essentially processes for neural network calculations. Preferably, the light location engine neural network application is defined as Neural Network Lite that operates when the mobile device (10) provides estimation with basic features of neural network without involving the location server (21). Such basic features of the neural network include a predicting method for heavy calculation on the server and a light calculation by using a Structured Query Language for mobile device (10). Heavy calculation involves the step of comparing the current measured sample with the training samples that were optimized, whereas light calculation involves the step of comparing measured samples of cell ID and RXL with the previous data stored in the mobile device (10). The Collector Application (24) is used to collect the RXLs and receive configuration setting sent by the location server (21).

The QOP Mapper (25) is used to map the quality of positioning (QOP) level with the type of network service required such as GSM, GPS, 3G, HSDPA and Wi-Fi. For example, an emergency state requires high QOP, thus GPS service is required to achieve more accurate location coordinates.

FIG. 2(a) shows a flowchart of a method for locating a mobile device (10) according to an embodiment of the present invention. In step 100, relevant data such as Cell ID and RXLs are collected wherein the collected data is then processed for performing predictions and comparisons of the system performance. The collected data is used in General Regression Neural Network algorithm. Thereon, a neural network training pattern and optimization is performed as in step 200. In step 300, a location request is performed with the Quality of Positioning (QOP) Mapper (25). In step 400, the location request is performed with the telecommunication service provider (30). In step 500, the mobile device (10) with the Phone Listener Application (23) operates with the Neural Network Lite. In step 600, the method searches for location server (21) if there is no internet connection. In step 700, a server search call is performed if there is no internet connection and the coordinates of the last location of the mobile device (10) is not available in the database (22).

FIG. 2(b) shows a flowchart of a method for processing collected data such as Cell ID and RXLs as in step 100 of FIG. 2(a). Initially, the location server (21) sends a command to collect RXLs and configuration setting to the Collector Application (24) as in step 101. The configuration setting is the command setting in changing the type of network service from 2G to 3G, 3G to 2G, LTE to 2G, or to turn on the Wi-Fi as well as GPS in order to collect the signal strength or RXLs. In step 102, the Collector Application (24) changes network settings by turning on the GPS and records Cell ID and RXLs in a Training Table.

Table 1 and Table 2 show the examples of Training Table for collected data and validation format. The table is used to collect relevant data and parameters used in the General Regression Neural Network. Firstly, the collected value is stored in the training table format, [T] where output for [T] is [y] = [lat, long] as captured by the GPS during data collection. The data is collected for several seconds in a static and moving motion so that proximity coordinates and entire area coordinates are collected. In Table 1, the user Cell ID and the Neighbours' Cell ID are recorded as N1 till N8 for all network services from 2G (GSM) during idle mode (no voice or no activity) and different mode such as 3G and HSDPA cells. The Neighbours' Cell ID are collected and recorded for the purpose of sampling. In Table 2, the RXLs for the corresponding Cell ID in Table 1 are recorded at the columns provided for the related network services.

Time/Date Serving Cell ID N1 N2 N3 N4 N5 N6 N7 N8

GSM (Idle) 10025

GSM (Voice) 10022

3G (Idle) 345

3G (Voice) 234

GPRS 10300

3G Data 340

HSDPA 540

Wi-Fi Mimos Guest TM N1 NN

Table 1

Table 2

Thereon, the recorded Cell ID and RXLs are stored in the database (22) by the Collector Application (24) and sent to the location server (21) within pre- configured time interval as in step 103. When the maximum sample size (N) set by the location server (21) is achieved, the Collector Application (24) turns off the GPS as in step 104.

FIG. 2(c) shows a flowchart of a method for processing a neural network training pattern and optimization as in step 200 of FIG. 2(a). Based on the collected data from the Training Table, the collected data is arranged based on Location Area Code (LAC) accordingly as in step 201. The RXLs are arranged based on the Cell IDs, network services, GPS coordinates, time and date as in step 202. In step 203, 80% of N samples of the collected data are used as Training Samples and 20% of N samples from the collected data are used as Estimator Samples. Preferably, the percentage of the sample is predefined based on the collected data. Training samples are samples used for training the neural network algorithm to derive a pattern for that particular location area. As certain location provides different Cell ID and RXLs, than other locations, the estimation or prediction of the proximity location of the mobile device (10) can be obtained. For example, if a grid of area is taken as 50 meter by 50 meter, the grid has different combination of RXLs and Cell ID as compared to the neighbouring grids (which are considered as different location). Thus, it is possible to estimate which grid area or location based on the RXLs and Cell IDs of the mobile device (10) and the accuracy of estimation increases if there is more than one RXLs and Cell IDs of various network types are involved in the estimation. On the other hand, Estimator samples are used to re-evaluate and test robustness of the trained neural network algorithm.

From the collected data, a smooth parameter (σ) is determined based on a General Regression Neural Network (GRNN). GRNN is used to train the ΓΤ] with [y], wherein the GRNN is defined as GRNN (σ, T, y). Thereon, another set of data of [T] is collected and known as [X]. Next, smooth parameter (σ) is tuned for GRNN (σ, T, y) in order to get a new [y(x)] output coordinate(s). Then, use Euclidean Distance to compare the initial collected [y(T)] to current [y(x)]'s error distance.

The collected data is used to plot a Cumulative Distribution Function (CDF) graph. Fundamentally, the CDF graph describes a probability that a real-valued random variable (X) with a given probability distribution will be found at a value less than or equal to (X). The distance between two points in the CDF graph is measured by using a Euclidean Distance method to find the difference of the location coordinates as in step 204. Thereon, the CDF of Location Estimate is computed to analyze the difference of the location coordinates. The Location Estimate refers to an estimation of the location of the mobile device (10) with respect to the actual location request. A benchmark is obtained in order to compare with the smooth parameter (σ). Preferably, the benchmark refers to the Federal Communications Commission USA (FCC) benchmark, wherein 67% of Estimator Samples from the CDF graph has less than 100m location accuracy and 95% of Estimator Samples from the CDF graph has less than 300m location accuracy as in step 205. Preferably, iteration value is obtained from the repetition calculation of the location estimate. In decision 206, if the smooth parameter (σ) is not within the benchmark or the difference (Diff **) between the first iteration value and the second iteration value is not too small, the method returns to step 101 of FIG. 2(a). If the smooth parameter (σ) is within the benchmark or the difference (Diff **) between the first iteration value and the second iteration value is too small, new correction factor (p) is calculated wherein for initial case, correction factor (ε) is assumed to be equal to 1 as in step 207. Thereon, the value of p is stored in the database (22) as in step 208. The value of p is used in the following Location Area Code (LAC) for obtaining actual samples of RXLs and Cell IDs without the GPS assist for searching the location. FIG. 2(d) shows a flowchart of a method for performing a location request with the QOP Mapper (25) as in step 300 of FIG. 2(a). The QOP Mapper (25) maps the QOP level with the type of network services required. The location server (21) sends the configuration setting to change the network service based on the QOP level required as well as the location request to the Phone Listener Application (23) for estimating the current location of the mobile device (10) as in step 301. The Phone Listener Application (23) starts the measurement process of the Cell ID and RXLs based on the QOP level required if the mobile device (10) has internet connection as in step 302. In decision 303, if the QOP level is low, the Cell ID, RXLs and/or Wi-Fi information are obtained as in step 304. Preferably, the Wi-Fi information is obtained if the mobile device is in an indoor area, wherein the Wi-Fi coordinates is used as the proximity location. Based on the Cell ID, RXLs and/or Wi- Fi information, the location server (21) translates the Cell ID coordinates or the Wi-Fi coordinates to location coordinates as in step 305. Then, the location server (21) translates the Wi-Fi coordinates to latitude and longitude coordinates in a map such as Google Map or Geographic Information System (GIS) as in step 314. If the QOP level is medium, the RXLs of that particular network service and Wi-Fi network service are obtained and sent to the location server (21) as in step 306. Thereon, current information is matched with the stored RXLs using a Statistical Method and Free Space Propagation Loss Formula, wherein the RXLs are converted into distance as in step 307. Statistical method includes the maximum possibility, average of the Cell ID obtained with the same RXLs and the Cell ID available in the database (22) to estimate the best coordinates for the mobile device (10). Free Space Propagation Loss Formula stated that Free-space path loss is proportional to the square of the distance between the transmitter and receiver, and also proportional to the square of the frequency of the radio signal. In step 308, at least three base stations coordinates are triangulated to estimate the location coordinates. Then, the location server (21) translates the obtained coordinates to latitude and longitude coordinates in a map such as Google Map or Geographic Information System (GIS) as in step 314.

If the QOP level is high, GPS is used to assist the location request. In decision 309, if the GPS is turned on, the location coordinates is directly obtained as in step 310. Then, the location server (21) translates the obtained coordinates in the map such as Google Map or Geographic Information System (GIS) as in step 314. If the GPS is turned off, Cell IDs and RXLs of the networks setting are obtained by changing the network services as shown in Table 1 and then sent to the location server (21) as in step 311. Based on the collected data, value of p, and GPS coordinates of the Cell IDs that have been processed by the neural network are used in the GRNN to estimate the current location of the mobile device (10) as in step 312. Later on, the General Regression Neural Network is used to estimate the current location by comparing the new Cell IDs or RXLs with the trained Cell IDs or RXLs as in step 313. Then the location server (21) translates the obtained coordinates to latitude and longitude coordinates in a map such as Google Map or Geographic Information System (GIS) as in step 314.

FIG. 2(e) shows a flowchart of a method for performing a location request with the telecommunication service provider (30) as in step 400 of FIG. 2(a). In step 401 , the location server (21) sends the enterprise phone number of the mobile device (10) that needs to be located for emergency, public safety or phone loss to the telecommunication service provider (30). The telecommunication service provider (30) uses its Mobile Location Centre to search for the location of the mobile device (10) as in step 402. Thereon, the telecommunication service provider (30) sends the coordinates of the mobile device (10) to the location server (21) if the SIM card is attached to the phone or the telecommunication service provider (30) sends the last active location of the mobile device (10) as in step 403. The location server (21) translates the received coordinates to latitude and longitude coordinates in a map such as Google Map or Graphic Information System (GIS) as in step 404. FIG. 2(f) shows a flowchart of a method for operating the mobile device (10) with the Phone Listener Application (23) which operates with the Neural Network Lite as in step 500 of FIG. 2(a). Initially, the Phone Listener Application (23) monitors battery and connectivity status of the mobile device (10). The Phone Listener Application (23) is operating with the Neural Network Lite, wherein the Neural Network Lite compares the stored Cell IDs with the Wi-Fi IDs within the enterprise or compares the stored Cell IDs with frequent travelled areas if there is no internet connection as in step 501. In decision 502, if there is no internet connection, Cell ID and Wi-Fi IDs, the method ends. If there is an internet connection, Cell ID and Wi-Fi IDs, the mobile device (10) locks the phone immediately or within a pre-configured time interval as in step 503. All access to the application of the mobile device (10) is denied and contents are wiped within a pre-configured time interval accordingly. From the heterogeneous networks of RXLs, the location server (21) uses any QOP level to estimate the coordinates wherein the initial default of QOP level is high as in step 504. A password or pin code is required to unlock the mobile device (10). Later on, a message is sent to the administrator to inform the unattended mobile device's coordinates as in step 505 if the Phone Listener (23) application in the mobile device (10) is not able to do the server search.

FIG. 2(g) shows a flowchart of a method for searching the location server (21) if there is no internet connection as in step 600 of FIG. 2(a). Firstly, if the location server (21) detects no active communication from the last location of the mobile device (10) for several hours or within a preconfigured time interval, the location server (21) sends a message or performs a voice call as in step 601. The Phone Listener Application (23) intercepts the message or voice call from the enterprise number and reads request command such as wipe, lock or location search with High QOP level wherein the mobile device (10) is assumed that it never undergoes any factory reset as in step 602. The Cell IDs and RXLs of the networks settings are obtained by changing networks and thereon, the enterprise message number are sent to the location server (21) as multiple message characters as in step 603 wherein the enterprise message number is a pre-stored phone number connected to the enterprise location server. The location server (21) reads the messages wherein the messages are extracted to proceed with location search using the General Regression Neural Network for the current LAC as in step 604. In decision 605, if there is no trained data for the LAC, Cell ID coordinates are used to find the proximity coordinates of the mobile device (10) as in step 606. Then, the location server (21) translates the received coordinates in the map such as Google Map or Graphic Information System (GIS) as in step 607. If there is trained data for the Location Area Code, the collected data, value of p and stored GPS coordinates of Cell IDs and RXLs are loaded from the LAC and the Cell IDs as in step 608. Later on, the General Regression Neural Network is used to estimate the current location by comparing the new Cell IDs or RXLs with the Cell IDs or RXLs as in step 609. Then, the location server (21) translates the received coordinates to latitude and longitude coordinates in a map such as Google Map or Graphic Information System (GIS) as in step 607.

FIG. 2(h) shows a flowchart of performing a server search call if there is no internet connection and the coordinates of the last location of the mobile device (10) is not available in the database (22) as in step 700 of FIG. 2 (a). Firstly, if the location server (21) detects no active communication from the last location of the mobile device (10) within a pre-configured time interval, the location server (21) performs a voice call as in step 701. If the voice call is answered, a recording is played to request the returning of the mobile device (10) to the address sent by the location server (21) via message. The Phone Listener Application (23) intercepts the voice call from the enterprise number regardless if the mobile device (10) is answered or not wherein the mobile device (10) is assumed that it never undergoes any factory reset as in step 702. In step 703, if the phone call or message of the mobile device (10) provides no response to the location server (21), the telecommunication service provider (30) identifies the location of the mobile device (10) based on International Mobile Station Equipment Identity (IMEI) of the mobile device (10). If the GPS cannot be activated to provide location coordinates, the Cell IDs and RXLs of the network settings are obtained by changing networks and the enterprise message number are sent to the location server (21) as multiple message characters as in step 704. In decision 705, if the message is successfully sent, the location server (21) reads the messages wherein the text messages are extracted to proceed with location search using the General Regression Neural Network for the current LAC as in step 706. If the message is not successfully sent, the method determines whether there is trained data in the Neural Network Lite as in decision 707. If there is no trained data in the Neural Network Lite, text parameters such as the Cell ID, LAC, IMEI and RXLs are provided to a text-to-speech component for converting the text parameters to speech as in step 708. If there is trained data in the Neural Network Lite, the trained data of the Neural Network Lite, value of p and stored GPS coordinates of trained Cell IDs and RXLs are loaded from the LAC and the Cell IDs as in step 709. In step 710, the Neural Network Lite is used to predict location coordinates if the GPS is not activated. Thereon, text-to-speech program is used to speak the text and IMEI repeatedly as in step 711. In step 712, the Emergency Number is called and the text to speech voice result is voice read. Later on, the emergency centre personnel takes action and utilises the telecommunication service provider (30) to locate the mobile device (10) based on the IMEI when the emergency centre personnel received the voice reading as in step 712 and 713. In particular, the user can utilize the location search server connected to the telecommunication service provider (30) to track the lost IMEI. While embodiments of the invention have been illustrated and described, it is not intended that these embodiments illustrate and describe all possible forms of the invention. Rather, the words used in the specifications are words of description rather than limitation and various changes may be made without departing from the scope of the invention.

Claims

1. A system for locating a mobile device comprises of:

a) at least a mobile device (10),

b) a mobile enterprise platform (20),

c) at least one telecommunication service provider (30);

wherein the mobile enterprise platform (20) is characterised in that it further includes:

i. a location server (21) for locating the mobile device (10) in different types of networks;

ii. a database (22) to store collected data including Cell Identification (ID), receive signal levels (RXLs) and the location coordinates of the mobile device (10);

iii. a Phone Listener Application (23) to respond to request from the location server (21) and collect RXL of the mobile device (10);

iv. a Collector Application to collect RXLs and configuration setting sent by the location server (21); and v. a Quality of Positioning (QOP) Mapper (25) to map the quality of positioning (QOP) level with the type of network required.

2. The system (100) as claimed in claim 1 , wherein the Phone Listener Application (23) includes a location engine application to provide estimation without involving the server of the location server (21).

3. A method for locating a mobile device (10) is characterised by the steps of: a. ) collecting and processing data including Cell Identification (ID), signal levels (RXLs) and location coordinates of the mobile device (10); b. ) performing a neural network training pattern and optimization;

c. ) performing a location request with a Quality of Positioning (QOP) Mapper

(25);

d. ) performing a location request with the telecommunication service provider

(30);

e. ) operating the mobile device (10) with the Phone Listener Application (23); f. ) searching for location server (21); and g.) performing a server search call if there is no internet connection and the coordinates of a last location of the mobile device (10) is not available in the database (22).

The method as claimed in claim 3, wherein collecting and processing data of the mobile device (10) includes the steps of:

a. ) sending a command to the Collector Application (24) for collecting RXLs and configuration setting by the location server (21);

b. ) changing network settings by turning on the GPS;

c. ) recording Cell ID and RXLs in a Training Table by the Collector

Application (24);

d. ) sending the recorded Cell ID and RXLs to the location server (21) within pre-configured time interval; and

e. ) turning off the GPS by the Collector Application (24) when a maximum sample size (N) set by the location server (21) is achieved.

The method as claimed in claim 3, wherein processing a neural network training pattern and optimization includes the steps of:

a. ) arranging the collected data based on Location Area Code (LAC);

b. ) arranging the RXLs based on the Cell IDs, network services, GPS coordinates, time and date;

c. ) training neural network algorithm to derive a pattern for a particular location area based on a predefined percentage of samples of the collected data;

d. ) testing the neural network algorithm using a predefined percentage of samples of the collected data;

e. ) determining a smooth parameter (σ) based on a General Regression

Neural Network (GRNN);

f. ) plotting a Cumulative Distribution Function (CDF) graph by using the collected data;

g. ) measuring the distance between two points in the CDF graph to find the difference of the location coordinates;

h. ) comparing the smooth parameter (σ) with a benchmark; i.) calculating a new correction factor (p) if the smooth parameter (σ) is within the benchmark or the difference (Diff **) between the first iteration value and the second iteration value is too small; and

j.) storing the value of p in the database (22).

The method as claimed in claim 3, wherein performing a location request with the QOP Mapper (25) includes the steps of:

a) mapping a Quality of Positioning (QOP) level with the type of network services required by the QOP Mapper (25);

b) sending the configuration setting to change the network service based on the QOP level required and the location request to the Phone Listener Application (23) by the location server (21);

c) measuring the Cell ID and RXLs based on the QOP level required by the Phone Listener Application (23);

d) if the QOP level is low, obtaining the Cell ID, RXLs and Wi-Fi information, and translating the Cell ID coordinates to a location coordinates by the location server (21);

e) if the QOP level is medium, obtaining the RXLs of that particular network service and Wi-Fi network service, sending the RXLs of that particular network service and Wi-Fi network service to the location server (21), matching the RXLs of that particular network service and Wi-Fi network service with the stored RXLs using a Statistical Method and Free Space Propagation Loss Formula, and triangulating at least three base stations coordinates to estimate the location coordinates;

a) if the QOP level is high, obtaining Cell IDs and RXLs of the networks setting by changing the network services, sending the Cell IDs and RXLs of the networks setting to the location server (21), loading the value of p, and GPS coordinates of the Cell IDs that have been processed by the neural network based on the collected data to estimate the current location of the mobile device (10), and estimating the current location by comparing the new Cell IDs or RXLs with the trained Cell IDs or RXLs; f) translating the Wi-Fi coordinates to latitude and longitude coordinates in a map by the location server (21). The method as claimed in claim 3, wherein performing a location request with the telecommunication service provider (30) includes the steps of:

a) sending the enterprise phone number of the mobile device (10) to the telecommunication service provider (30) by the location server (21);

b) searching for the mobile device (10) by using a Mobile Location Centre by the telecommunication service provider (30);

c) sending the coordinates of the mobile device (10) to the location server (21) if the SIM card is installed in the mobile device (10);

d) sending the last active location of the mobile device (10) by the telecommunication service provider (30); and

e) translating the location coordinates to latitude and longitude coordinates in a map by the location server (21).

The method as claimed in claim 3, wherein operating the mobile device (10) with the Phone Listener Application (23) includes the steps of:

a) monitoring battery and connectivity status of the mobile device (10) by the Phone Listener Application (23);

b) comparing the stored Cell IDs with the Wi-Fi IDs within the enterprise or comparing the stored Cell IDs with frequent travelled areas if there is no internet connection;

c) locking the mobile device (10) within a pre-configured time interval if there is internet connection, Cell ID and Wi-Fi IDs;

d) estimating the coordinates from the heterogeneous networks of RXLs by using any QOP level by the location server (21); and

e) sending the coordinates of the mobile device (10) to the administrator if the Phone Listener Application (23) in the mobile device (10) is not able to do the server search.

The method as claimed in claim 3, wherein searching the location server (21) includes the steps of:

a) sending a message or performing a voice call by the location server (21) if the location server (21) detects no active communication from the last location of the mobile device (10) within a preconfigured time interval; b) intercepting the message or voice call from the enterprise number and reading request command with high QOP level by the Phone Listener Application (23);

c) obtaining the Cell IDs and RXLs of the networks settings by changing networks;

d) sending the enterprise message number to the location server (21), wherein the enterprise message number is a pre-stored phone number connected to the enterprise location server;

e) reading the messages by the location server (21), wherein the messages are extracted to proceed with location search using the General Regression Neural Network for the current Location Area Code (LAC); f) determining the proximity coordinates of the mobile device (10) by using the Cell ID coordinates if there is no trained data for the LAC; and g) translating the location coordinates to latitude and longitude coordinates in a map by the location server (21).

10. The method as claimed in claim 3, wherein performing a server search call if there is no internet connection and the coordinates of the last location of the mobile device (10) is not available in the database (22) includes the steps of: a) performing a voice call by the location server (21) if the location server (21) detects no active communication from the last location of the mobile device (10) within a pre-configured time interval;

b) playing a recording to request the returning of the mobile device (10) to the address sent by the location server (21) via message if the voice call is answered;

c) intercepting the voice call from the enterprise number by the Phone Listener Application (23);

d) identifying the location of the mobile device (10) based on International Mobile Station Equipment Identity (IMEI) of the mobile device (10) by the telecommunication service provider (30) if the phone call or message of the mobile device (10) provides no response to the location server (21); e) obtaining the Cell IDs and RXLs of the network settings by changing networks if the GPS cannot be activated to provide location coordinates; f) sending the enterprise message number to the location server (21); g) reading the messages by the location server (21) if the message is successfully sent, wherein the text messages are extracted to proceed with location search using the General Regression Neural Network for the current LAC;

h) determining whether there is trained data in the Neural Network Lite if the message is not successfully sent;

i) providing the Cell ID, LAC, IMEI and RXLs information if there is no trained data in the Neural Network Lite;

j) loading the trained stored data of the Neural Network Lite, value of p and stored GPS coordinates of trained Cell IDs and RXLs from the LAC and the Cell IDs if there is trained data in the Neural Network Lite;

k) predicting the location coordinates by the Neural Network Lite if the GPS is not activated; and

I) sending the Cell IDs and IMEI number of the mobile device (10) to an emergency number.